2709
J . Phys. Chem. 1986, 90, 2709-2715
Conversion of Acetylene to Benzene over Palladium Single-Crystal Surfaces. 2. The Effect of Additives M. A. Logan,+T. G. Rucker, T. M. Gentle,*E. L. Muetterties,s and G. A. Somorjai* Materials and Molecular Research Division, Lawrence Berkeley Laboratory and Department of Chemistry, University of California, Berkeley, Berkeley, California 94720 (Received: October 11, 1985)
The conversion of acetylene to benzene over palladium single-crystal surfaces (( 11l), (loo), and (1 10)) modified with submonolayer quantities of silicon, phosphorus, sulfur, chlorine, and potassium has been studied under both ultrahigh vacuum and atmospheric pressure conditions. At atmospheric pressures electron-donatingadditives enhance the rate of benzene formation and electron-acceptingadditives reduce the rate. Potassium and silicon reduce the amount of surface carbon that accumulates during the reaction and thereby increase the rate of benzene formation relative to the undoped surface. The general trends of the (1 1l), (loo), and (1 10) faces indicate that at low pressures potassium suppresses, silicon and phosphorus enhance, and sulfur and chlorine leave unchanged the benzene yield. Silicon on the (100) face showed the greatest enhancement in the benzene yield possibly due to surface compound formation.
1.0. Introduction Additives are widely used in both low-pressure stoichiometric reactions and high-pressure catalytic reactions to alter the product distribution, to raise the yield, or to increase catalyst lifetime. It is not well understood how the additives modify the catalyst surface for these reactions. Two major types of interactions are possible: (1) a geometric effect due to site blocking and (2) an electronic effect due to the donation or depletion of electron density in the near surface region. The conversion of acetylene to benzene has been shown to occur both in ultrahigh vacuum and at higher pressures on a variety of palladium surfaces with different morphologies (single crystals,'+ films,' and small metal particles supported on alumina') as well as over many homogeneous catalysts.'s12 A comparison of the low-pressure and high-pressure reaction on palladium single crystals is described in detail in the previous paper.' In this paper we study the effects of silicon, phosphorus, sulfur, chlorine, and potassium on the chemisorptive and catalytic properties of palladium (1 l l ) , (loo), and (1 10) single-crystal surfaces on the cyclotrimerization of acetylene. The differences in additive behavior in the low-pressure and the high-pressure (1 atm) reactions are compared. The effect of the additives is quantified by catalytic rates of benzene production from acetylene, temperature-programmed desorption (TPD) of hydrogen, acetylene, ethylene, benzene, and carbon monoxide, and by work function measurements. 2.0. Experimental Section Experiments were carried out in two ultrahigh vacuum (UHV) chambers. . Low-pressure experiments were conducted in a standard UHV chamber with a base pressure of 1 X Torr. The chamber was equipped with a four-grid retarding field analyzer and glancing incidence electron gun for Auger electron spectroscopy ( A B ) and for work function changes as the surface was modified by additives. Changes in the work function were measured by the change in the onset energy of the secondary electron cascade and were reproducible to within *lo%. The high-pressure experiments were performed in a UHV/ high-pressure apparatus designed for combined UHV surface analysis and high-pressure reaction studiesI3 using small surface area catalyst samples. This chamber was equipped with four-grid electron optics for low-energy electron diffraction (LEED), a double pass cylindrical mirror analyzer with coaxial electron gun for A m , an Ar+ sputtering gun for crystal cleaning, a quadrupole mass spectrometer, and a retractable internal isolation cell that constitutes part of a microbatch reactor operating in the pressure 'Present address: J. C. Schumacher Co., P.O. Box 158, Oceanside, CA 92054. *Present address: Dow Corning Corp., Midland, MI 48640. #Deceased January 1984.
0022-3654/86/2090-2709$01.50/0
range of 10-2-20 atm. The experimental procedure for both the high-pressure and the low-pressure studies are described in detail in the previous paper.' For additive-doping studies, coverages were estimated from the relative Auger sensitivities of the most intense peaks,I4 which is accurate only for low coverages. The relative additive coverage is given as OA = (hASA)/(hASA h&), where hA and hs are the peak heights of the adsorbate and substrate and S A and Ss are the Auger sensitivity factors for the adsorbate and substrate. Potassium was deposited under vacuum from a SAES getters mounted 2 cm from the sample. Silicon (Matheson grade SiH4), phosphorus (Matheson grade PH3), sulfur (Matheson grade H2S), and chlorine (Matheson H.P. grade C12) were deposited on the crystal surface at 300 K through a variable rate leak valve. Subsequently, the crystal surface was heated (700 K) to desorb HZ,and the coverage was determined by A B . Additive coverages remained unchanged, as observed with AES during both UHV and high-pressure experiments. Product formation was followed by gas chromatography. Initial reaction rates were determined graphically from the slopes of product accumulation curves as a function of time over the first 2-3 h of reaction time and were reproducible to within *lo%. Blank experiments performed on Pd covered with graphitic carbon (formed by heating the crystal in a hydrocarbon atmosphere at 750 K) showed only a low level of catalytic activity. For example, at 575 K the activity of the carbon-covered surface was less than 10% of the activity of clean palladium at the lowest reaction temperature (275 K) used in this study.
+
3.0. Results The cyclotrimerization of acetylene was observed on adsorbate-modified (1 1l), (loo), and (1 10) single crystals of Pd. These (1) Rucker, T. G.; Logan, M. A,; Gentle, T. M.; Muetterties, E. L.; Somorjai, G. A. J . Phys. Chem., preceding article in this issue. (2) Tysoe, W. T.; Nyberg, G. L.; Lambert, R. M. J. Chem. SOC.,Chem. Commun. 1983, 623. ( 3 ) Tysoe, W. T.; Nyberg, G. L.; Lambert, R. M. Surf. Sic. 1983, 135, 128. (4) Gentle, T. M.; Muetterties, E. L. J . Phys. Chem. 1983, 87, 2469. ( 5 ) Sesselmann, W.; Woratschek, B.;Ertl, G.; Kuppers, J. Surf: Sic. 1983, 130, 245. (6) Gentle, T. M.; Grassian, V. H.; Klarup, D. G.; Muetterties, E. L. J . Am. Chem. SOC.1983, 105, 6166. (7) Gates, J. A.; Kesmodel, L.L. Surf. Sci. 1983, 124, 6 8 . (8) Gates, J. A.; Kesmodel, L. L. J . Chem. Phys. 1982, 76, 4281. (9) Kesmodel, L. L.; Waddill, G. D.; Gates, J. A. Surf: Sci. 1984, 138,464. (10) Parshall, G. W. Homogeneous Catalysis; Wiley: New York, 1980; p 165. (11) Maitlis, P. M. Acc. Chem. Res. 1976, 9, 93. (12) Vollhardt, K. P. C. Angew. Chem. 1984, 123, 539. (13) Chambrey, M.; Cabrera, A. L.; Spencer, N. D.; Kozak, E.; Davies, P. N.; Somorjai, G. A. Rev. Sci. Znstrum. 1982, 53, 1893. (14) (a) Palmberg, P. W.; Riach, G. E.; Weber, R. E.; MacDonald, N. C. Handbook of Auger Electron Spectroscopy; Physical Electronics Ind. Inc.: Edina, MN, 1972. (b) Mroczkowski, Lichtman, Surf. Sci. 1983, 131. 159.
0 1986 American Chemical Society
2710 The Journal of Physical Chemistry, Vol. 90, No. 12, 1986
1 I
Logan et al.
Pd (11l)-C,H, I \
I
250
470
Temperature (K)
Temperature (K) Figure 1. The TPD spectra of acetylene dosed on clean Pd(ll1) (6 langmuirs of acetylene, 130 K) are displayed. The four desorption
products are benzene from cyclotrimerization,ethylene from hydrogenation, acetylene from reversible molecular desorption, and hydrogen from decomposition.
L
.-cn v)
I
I
150
250
475
600
Temperature (K) Figure 2. The benzene TPD from 6 langmuirs of acetylene dosed on additive-covered covered Pd( 111) at 130 K is displayed. The high-temperature peak shifts to higher temperatures: from 470 K on the silicon-dosed surface to 490 K on the chlorine-covered surface. The phosphorus-covered surface leads to a thermal desorption trace for benzene with four maxima. All spectra magnified by a factor of 2.
surfaces have been studied in ultrahigh vacuum by TPD and a t 1 atm pressure by the high-pressure/low-pressure apparatus to obtain kinetic information on the reaction. Detailed information on the trimerization reaction in both pressure regions on clean Pd is reported in the previous paper.' 3.1. Effect of Adatoms ( S i , P, S, Cl, and K ) on the LowPressure Stoichiometric Reaction over Palladium Single Crystals. Thermal desorption studies were carried out in UHV for three Pd surfaces (( 11l ) , (loo), and (1 10)) to determine the effect of adatoms (Si, P, S, C1, and K) on the acetylene cyclotrimerization, the decomposition of acetylene to carbon and hydrogen, the self-hydrogenation to ethylene, and the reversible molecular desorption of acetylene. All doses were 6 langmuirs (1 langmuir = 1 X 10" Torr s) and at 130 K unless stated otherwise. 3.1.1. P d ( l l 1 ) . As reported previously' TPD of 6 langmuirs of acetylene from clean P d ( l l 1 ) resulted in some reversible molecular desorption a t low temperatures (190 K), self-hydrogenation to form ethylene (300 K), decomposition to form hydrogen (430-830 K), and formation of benzene (250-490 K) (Figure 1). Adsorption of silicon (0 = 0.25) on Pd( 111) did not significantly perturb the benzene formation as shown by the thermal desorption results (Figure 2). The only effect of silicon was to slightly decrease acetylene decomposition and enhance the high-temperature desorption peak of benzene. Phosphorus, on the other hand, greatly suppressed the acetylene decomposition, increased
Figure 3. Benzene TPD spectra are compared after dosing silicon-doped Pd( 11 1) with 4 langmuirs of acetylene or benzene at 130 K. Benzene formed from acetylene yields a broader high-temperaturepeak. Higher acetylene exposures enhance the high-temperaturepeak, whereas larger benzene exposures increase the low-temperature peak.
the yield of ethylene, and gave rise to a thermal desorption spectrum of benzene (from the cyclotrimerization reaction) with maxima at 200, 300, 400, and 470 K, while doubling the benzene yield (Figure 2). Sulfur adatoms also affected the thermal behavior of chemisorbed acetylene (Figure 2). At low sulfur coverages (0 < 0.2), the decomposition to hydrogen and the ethylene yield decreased, while the benzene yield increased significantly. The low-temperature benzene peak intensity almost doubled relative to the clean surface and a new peak appeared at 430 K. At higher sulfur coverages (0 < 0.25), the formation of ethylene was fully suppressed and the decomposition to Hz was very low. At these high sulfur coverages, the benzene desorption spectrum comprised a broad peak with a single maximum at 260 K equal in intensity to the clean surface. Like phosphorus and sulfur, chlorine adatoms also enhanced benzene formation (Figure 2). In contrast to sulfur, the benzene desorbed largely in the high-temperature regions, with maxima at 415 and 490 K, although there was a smaller peak at 270 K. Chlorine also suppressed ethylene formation and acetylene decomposition. The adatoms increased acetylene cyclotrimerization and decreased side reactions in order from silicon to phosphorus to sulfur on Pd( 111). Chlorine was not as effective as sulfur, which almost fully suppressed all acetylene reactions except for cyclotrimerization. Thus Pd( 1 11)-S is a selective catalyst for acetylene cyclotrimerization under UHV conditions. The main difference in the TPD spectra between the Pd( 11 1)-S and Pd( 11 1)-C1 surfaces was that the Pd(1 11)-S surface promoted primarily the low-temperature cyclotrimerization reaction and the Pd( 11 1)-C1 surface promoted the high-temperature process. TPD studies of adsorbed benzene on adatom-covered palladium surfaces yield similar desorption spectra as benzene formed from acetylene. Figure 3 shows the effect of silicon (0 = 0.25) on Pd( 11 1) on the desorption of benzene and benzene formed from acetylene. Silicon enhanced the high-temperature TPD peak of benzene relative to the clean surface for both acetylene and benzene adsorption. Higher exposures of both acetylene and benzene led to an enhancement of the high-temperature (470 K) benzene desorption peak. 3.1.2. P d ( l l 0 ) . Acetylene chemisorbed on clean Pd( 110) resulted in reversible molecular desorption, decomposition, selfhydrogenation, and cyclotrimerization. The yield of benzene from the clean Pd( 110) crystal face was found to be approximately one-sixth that of the clean Pd( 11 1) crystal face. Silicon at low surface coverages (0 < 0.25) enhanced the benzene yield through the high-temperature desorption peak (Figure 4) and decreased acetylene decomposition. At BSi = 0.37 the yield of benzene increased by a factor of 3.5 (relative to the clean surface), while the ethylene and hydrogen yields were less than that observed for the clean Pd( 1 10) surface. Phosphorus on the (1 10) surface of palladium increased cyclotrimerization and self-hydrogenation. Benzene desorption was found to increase at phosphorus coverages greater than 0 = 0.2. Again, as on
The Journal of Physical Chemistry, Vol. 90, No. 12, I986 2711
Conversion of Acetylene to Benzene
u
-\A
W 250
0.25 ML S
(X1O)j
M
Clean Pd (110) (X10)
I
I
1
250
500
I
Temperature (K)
460
Temperature (K)
Figure 4. The benzene TPD spectra from additive-covered Pd( 1 IO) after a 6-langmuir exposure of acetylene at 130 K are displayed. Silicon and
phosphorus increase the high-temperature peak. Sulfur enhances the low-temperature benzene desorption maxima slightly and chlorine reduces the overall yield and broadens the two maxima into one peak. Pd(l11) and Pd(100), the gain in benzene yield was in the high-temperature desorption peak (Figure 4). At higher phosphorus coverages, two benzene TPD maxima were observed at 270 and 440 K. The high-temperature maximum was increased by as much as a factor of 4 ($ = 0.42) relative to the clean surface, while the ethylene yield remained at the level of the low coverage surface. The amount of hydrogen desorption decreased. The benzene yield increased slightly, the ethylene yield decreased, and the molecular desorption of acetylene increased with increasing amounts of sulfur on the surface. At higher sulfur coverages (0, = 0.4) no ethylene was detected, very little hydrogen desorbed, and the amount of benzene formed was approximately the same as on the clean Pd( 110). Addition of chlorine to Pd( 110) decreased decomposition, hydrogenation, and cyclotrimerization of acetylene. At a chlorine coverage of 0 = 0.30, a very weak ethylene desorption peak was observed at 290 K and a broad benzene desorption maximum was observed at 370 K (Figure 4). 3.1.3. Pd(100). A TPD experiment after acetylene adsorption on clean Pd( 100) yielded reversible molecular desorption, decomposition, hydrogenation, and cyclotrimerization (Figure 5AD). The yield of benzene from the clean Pd(100) crystal face was one-twentieth that of the clean Pd( 1 11) crystal face. Potassium on the (100) face of palladium inhibited benzene formation (Figure SA). The yield of ethylene and hydrogen decreased by 95% (Figure 5, B and D), and the molecular desorption of acetylene was slightly reduced (20-25%) (Figure 5C). At 0 = 0.25 (or higher) coverages of potassium, no benzene formation was observed. Silicon on Pd( 100) substantially enhanced benzene and ethylene formation (Figure 5, A and B). The two benzene TPD peaks were similarly affected by silicon promotion on Pd( 100) and Pd( 111); that is, the high temperature peak increased and accounted for a sixfold increase in the yield of benzene at a silicon coverage of 0 = 0.10. Higher coverages of silicon on Pd(100) (e = 0.38) further increased the yield of benzene up to 20 times that of the clean surface (Figure 6). The ethylene yield was also enhanced by factor of 10. The decomposition reaction increased as well, although the hydrogen desorption remained constant. Phosphorus adsorbed on Pd( 100) influenced the chemistry of chemisorbed acetylene. The yield of benzene was enhanced, but not to the extent observed on silicon-covered Pd(100). At a phosphorus coverage of 0 = 0.17, two benzene maxima were observed, a t 350 and 520 K. Low coverages (e < 0.2) did not change the benzene or ethylene yields. At coverages of 0.3 < Bp < 0.5, the amount of both benzene and ethylene increased by a factor of 3 to 5 (Figure 5, A and B). With increasing phosphorus coverages the benzene desorption maxima were also observed to shift to higher temperatures, primarily to the high-temperature desorption peak at 480 K.
0.25 ML S (X4) 0.30 ML P (X5)
0.25 ML K (X10)
305
Temperature (K)
I /I%
I f I\
0.15 ML CI
I
0.25 ML S
0.30ML P I
ti
-
0.25 ML Si
\
4-
a-
0.25 ML K
clean Pd( 100) x 1 rn
9 ,190 ,
310
Temperature (K)
1 0.25 ML S (Xl)
0.25 ML K (X10)
4-
Clean Pd
(1OO)(X11 445
625
Temperature (K) Figure 5. Parts A-D (from top to bottom) present the desorption traces
for benzene, ethylene, acetylene, and hydrogen from additive-covered Pd(100) dosed with 6 langmuirs of acetylene at 130 K. Note the scaling factors. The benzene desorption changes both in the desorption temperature and desorption yield. Potassium suppresses desorption of all products except for acetylene. Silicon-coveredPd( 100) produces 15 times more benzene than the clean surface.
2712
The Journal of Physical Chemistry, Vol. 90, No. 12, 1986
Logan et al.
6 langmuir dose
x 3
a
.=
a c
Q 7
;f
8-
i 0 Si
M CI P
II
m-=
25
IS
j
N2 850 Torr
-
~
‘0
0.1
0.2
0.3
0.4
0.5
\
\
os1 .P
-
Additive coverage (monolayer) Figure 6. The benzene TPD peak area on Pd(100) is shown as a function of additive coverage after dosing with 6 langmuirs of acetylene at 130 K. Electron-withdrawing additives, sulfur and chlorine, and electron-
neutral phosphorus alter benzene formation to the same extent: initially increasing benzene formation and then for coverages greater than 0 = 0.22 decreasing benzene formation. Silicon enhances benzene formation to 0 = 0.38 (highest coverage studied) at which point 20 times more benzene forms than on the clean surface. Low sulfur coverages on Pd( 100) (8, < 0.25) decreased the yield of ethylene (Figure 5B) and only slightly increased the benzene yield (Figure SA). The desorption of hydrogen decreased markedly for all sulfur coverages, and the desorption spectra was very broad, ranging from 435-610 K (Figure 5D). The major change observed at higher sulfur coverages was the decrease in ethylene yield. For example, at 6s = 0.25 the ethylene yield was one-half that of the clean surface (Figure 5B). Chlorine was found to enhance the yield of both ethylene and benzene (Figure 5, A and B). The effect was much smaller than that observed for either silicon of phosphorus. For coverages of OCl = 0.3 the yield of benzene was 5 times higher than for the clean surface. The additive coverage dependence of benzene formation was studied on the Pd(100) surface (Figure 6). Sulfur and chlorine enhanced benzene formation up to coverages of 8 = 0.25 and at higher coverages hindered benzene formation. Silicon and phosphorus increase benzene production to the highest coverages studied (8 = 0.4). 3.2. Effect of Adatoms ( K , Si, P, S , and C1) on the HighPressure Catalytic Reaction on Palladium Single Crystals. High-pressure catalytic studies were carried out over three palladium surfaces ((I l l ) , (loo), and (1 10)) to determine the effect of adatoms on the acetylene cyclotrimerization. Reactions were carried out between 290 and 620 K at an acetylene pressure of 200 Torr. Reaction rates were calculated by assuming that each palladium surface atom was one reaction site, and no correction was made when adatoms were deposited on the surface. This leads us to report conservative values for the reaction rates. All reactions produced benzene and very small amounts of C4hydrocarbons. Small amounts of ethylene (
